1. Field of the Invention
The present invention relates to an image retrieval apparatus for extracting a frame as still image data from a plurality of frames included in moving image data, and a method thereof.
2. Description of the Related Art
In recent years, the spread of digital cameras, digital camcorders, and digital television broadcasts, and the increasing speed of Internet communications have facilitated the handling of moving image content. There are various kinds of moving image content, including scenic shots recorded with a digital camera and informational TV programs.
Consideration has also been given to technology for extracting and using images from moving images as still images. Japanese Patent Laid-Open No. 2005-236483 discloses technology for automatically selecting suitable frames as still images from moving images taken with a digital camera.
Specifically, consecutive frames in a moving image are compared with one another to determine whether the images are similar. If there is at least a prescribed number of consecutive frames that are similar, it is determined that these frames are suitable as still images, and selected frames are printed.
However, while the textual information has more importance than the images in the case of informational TV programs, cooking shows and the like, users are not able to extract frames displaying textual information as still images with the above conventional technology.
That is, in the case of cooking shows, for example, recipes are continuously displayed on the side of the screen for a given period of time, while the content and layout in other parts of the screen changes dynamically as the participants prepare food and the like.
In such cases, these frames will not be determined as suitable for still images when using the above conventional technology, because the whole screen does not remain static for a given period of time. That is, if users want to extract frames showing recipes displayed during a cooking show or various textual information presented during an informational program, they are faced with the hassle of fast-forwarding through the moving images and locating the desired frames themselves.